Villarreal FD, Das GK, Abid A, Kennedy IM, Kültz D

15 Pages·2014·2.22 MB·English

Sublethal Effects of CuO Nanoparticles on Mozambique
Tilapia (Oreochromismossambicus) Are Modulated by
Environmental Salinity
Fernando D. Villarreal1*., Gautom Kumar Das2., Aamir Abid2, Ian M. Kennedy2, Dietmar Ku¨ltz1
1DepartmentofAnimalScience,UniversityofCalifornia-Davis,Davis,California,UnitedStatesofAmerica,2DepartmentofMechanicalandAerospaceEngineering,
UniversityofCalifornia-Davis,Davis,California,UnitedStatesofAmerica
Abstract
Theincreasinguseofmanufacturednanoparticles(NP)indifferentapplicationshastriggeredtheneedtounderstandtheir
putativeecotoxicologicaleffectsintheenvironment.Copperoxidenanoparticles(CuONP)aretoxic,andinduceoxidative
stressandotherpathophysiologicalconditions.TheuniquepropertiesofNPcanchangedependingonthecharacteristicsof
themediatheyaresuspendedin,alteringtheimpactontheirtoxicitytoaquaticorganismsindifferentenvironments.Here,
Mozambiquetilapia(O.mossambicus)wereexposedtoflamesynthesizedCuONP(0.5and5mg?L21)intwoenvironmental
contexts:(a)constantfreshwater(FW)and(b)stepwiseincreaseinenvironmentalsalinity(SW).SublethaleffectsofCuONP
were monitored and used to dermine exposure endpoints. Fish exposed to 5mg?L21 CuO in SW showed an opercular
ventilationrateincrease,whereasfishexposedto5mg?L21inFWshowedamilderresponse.DifferenteffectsofCuONPon
antioxidant enzyme activities, accumulation of transcripts for metal-responsive genes, GSH:GSSG ratio, and Cu content in
fishgillandliveralsodemonstratethatadditiveosmoticstressmodulatesCuONPtoxicity.Weconcludethatthetoxicityof
CuONPdependsontheparticularenvironmentalcontextandthatsalinityisanimportantfactorformodulatingNPtoxicity
infish.
Citation:VillarrealFD,DasGK,AbidA,KennedyIM,Ku¨ltzD(2014)SublethalEffectsofCuONanoparticlesonMozambiqueTilapia(Oreochromismossambicus)Are
ModulatedbyEnvironmentalSalinity.PLoSONE9(2):e88723.doi:10.1371/journal.pone.0088723
Editor:StephenJ.Johnson,UniversityofKansas,UnitedStatesofAmerica
ReceivedOctober8,2013;AcceptedJanuary10,2014;PublishedFebruary10,2014
Copyright:(cid:2)2014Villarrealetal.Thisisanopen-accessarticledistributedunderthetermsoftheCreativeCommonsAttributionLicense,whichpermits
unrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalauthorandsourcearecredited.
Funding: Support for this study was provided by the National Institute of Environmental Health Sciences, Award Number P42ES004699 and W. M. Keck
Foundationresearchgrant.Thefundershadnoroleinstudydesign,datacollectionandanalysis,decisiontopublish,orpreparationofthemanuscript.
CompetingInterests:Theauthorshavedeclaredthatnocompetinginterestsexist.
*E-mail:[email protected]
.Theseauthorscontributedequallytothiswork.
Introduction 50%ofcells)rangingfrom3to31.07mg?L21inthemacrophage
celllineTHP-1[16].MuchofthetoxicologicalstudiesofCuONP
Manufactured nanoparticles (NP),with sizesranging from1 to
todatehasbeenfocusedonatmosphericorinhalationexposures.
100nm in two or three of its dimensions [1], have been
Incontrast,veryfewstudieshaveanalyzedCuONPeffectsonfish.
increasingly employed in a broad range of applications including
Oneofthemshowedthat,inFWcarp,CuONPwerenotacutely
environmental remediation, cosmetics, biomedicine, material
toxic, although growth was inhibited while CuO NP content in
sciences and electronics [2,3,4]. The unique physical properties
tissuesincreasedovertime,particularlyinintestine,gillsandliver
ofNParemainlyattributedtotheirhighsurfacetovolumeratio,
[17].Moreover,theauthors observeda reduction inthetailbeat
withalargeproportionoftheatomsbeingexposedonthesurface frequency after 96hours of exposure at 10mg?L21. Concentra-
compared to the bulk material [5]. In addition, factors including tionsaslowas0.5mg?L21ofCuONPinterferedwithhatchingof
chemical composition, size, shape, solubility and aggregation are
FWzebrafishembryos[14].CuONPtoxiceffectshavealsobeen
generally considered to be important parameters that determine
studied in otheraquatic organisms [18,19,20].
thepropertiesofNP[6].AmongmanufacturedNP,themetaland
TheexpandeduseofmanufacturedNPinagrowingnumberof
metaloxideNPareemployedinabroadrangeofapplications[7].
products and the potential for release of these materials into the
Copper (II) oxide (CuO) NP have been used to improve the
environmentisraisingconcernsabouttheirputativeenvironmen-
efficacy of gas sensors, as high-transition-temperature supercon-
tal toxicity [21]. Major technical limitations for assessing such
ductors, as catalysts, andas antimicrobial agents [8,9,10,11,12].
ecotoxicological impact include the difficulty of isolating manu-
Cu-basedNP,asothermetal-basedNP,couldposeabiological
factured NP from complex samples as well as their diverse
hazardriskbecauseoftheirabilitytoinducemetabolicalkalosis,to
chemical behaviors (aggregation, dissolution, surface coatings)
releasemetalionstothemedium(orintracellularly),andtoinduce
[3,22].Currently,dataontheconcentrationsorchemicalstatusof
the intracellular generation of reactive oxygen species (ROS)
NP being released to the environment are scarce [23,24]. As an
[13,14].CuONPshowedhighertoxicitythanthatofothermetal-
alternative for actual measurements, mathematical models have
containingNPtowardshumanepitheliallungcells[15],whilethey
been devised to determine the environmental concentration of
werefoundtodisplayTC (toxicconcentrationtoinducedeathof
50 manufacturedNP.Suchmodelssuggestthatconcentrationsinthe
PLOSONE | www.plosone.org 1 February2014 | Volume 9 | Issue 2 | e88723
CuONPEffectsonFishatDifferentSalinities
order of ng?L21 or mg?L21 could already be present in certain CuO NP characterization
environmental niches [25,26,27]. TransmissionElectronMicroscopy(TEM). TEMimages
In general, the toxicity of metallic and metal oxide NP is wereacquiredusingaPhillipsCM-12TEMoperatingat120kV.
attributed only partially to the release of ions resulting from A drop of NP dispersion was put onto a formver carbon film
dissolution.ThepropertiesofthesurfaceoftheNParesuspected supported on a 400mesh copper grid (3mm in diameter) and
to affect the induction of the toxicological response [6,21,28,29]. allowed to dry in air at room temperature. The grid was then
Inductionofoxidativestress,mediatedbyreactiveoxygenspecies mounted intothevacuum chamberforimaging.
(ROS) generation at the NP surface, has been suggested to be a Powder X-ray Diffraction (XRD). Approximately 100mg
major culprit in the toxicological effects of multiple metal-based of powder sample was stirred gently to break up lumps. The
NP [6,30,31,32]. Importantly, surface properties, particle aggre- powderysamplesoftheNPwerethenspreadevenlyontoazero-
gation status and dissolution attributes of NP are determined by backgroundholder.Step-scanX-raypowderdiffractiondatawere
the characteristics of the medium in which they are suspended, collected over the range of 2h range of 20–85u on a Scintag
including pH, ionic strength or the presence of biomolecules. powder x-ray diffractometer (XRD) with Cu Ka radiation
Therefore, the toxicity of NP towards aquatic organisms can be (l=1.54056 A˚, operated at 45kV and 40mA) with 4 mm
expectedtodependonwaterqualityparameters,includingsalinity divergence slit, 1 mm scattering slit, and 0.2mm receiving slit.
[28]. Thescanningstepsizewas0.015uin2hwithacountingtimeof1 s
Studies comparing the effects of NP in different aquatic per step.
environments are, however, scarce [33,34]. Nonetheless, copper Dynamic light scattering (DLS) and Zeta Potential (ZP)
ions toxicity in Fundulus heteroclitus is dependent on the environ- measurements. The hydrodynamic size distribution and state
mental salinity [35]. Additionally, several aquatic organisms ofagglomeration/dispersionoftheCuONPwasanalyzedbyDLS
showed an increased tolerance to copper ions with increasing usingaBrookhavenZetaPlusinstrument(BrookhavenInstrument
salinity[36].Similartrendswereobservedfornickel[37]andzinc Inc. Holtsville, NY). The effective diameter particle size distribu-
[38] inF.heteroclitus andKryptolebias marmoratus. tions were calculated by the 90Plus software (Brookhaven
Instruments). The effective diameter corresponds to the first
We have investigated the extent of toxicity of CuO NP on
cumulantofthecorrelationcurveandformultimodaldistributions
Mozambique tilapia (Oreochromis mossambicus) at different salinities
it is weighted roughly by scattering intensity. Samples (i.e. NP
by simulating a fresh water habitat and an estuarine habitat
dispersed in water in fish tanks; concentration 0.5 or 5mg?L21,
subject totidally induced salinity fluctuations.
pH,8.25) were taken from the tanks at every 24h for analysis.
Tilapiaarewidelydistributedinthewildinestuarineandinland
Samplesweretransferredintoa4 mlcuvettewithoutanyfiltration
habitats and extensively used in aquaculture [39]. Mozambique
to determine the size and assess the extent of aggregation. The
tilapiaisawell-establishedmodelforinvestigatingosmoregulatory
mean particle counts obtained in the measurements were 79 and
mechanismsbecauseofthisspecieshighcapacityforadaptingtoa
335kcps for 0.5 and 5 mg?L21 samples respectively. The ZP or
wide range of environmental salinities [40,41,42]. Furthermore,
overallsurfacechargeofNPsampledfromfreshwater(FW)tanks
tilapia has been used for many toxicological studies, including
wasdeterminedusingthesameZetaPlusinstrument(Brookhaven
studiesinvestigating theeffectsofmetals[43,44]andmetallicNP
Instruments Corp.).
[45,46].
Brunauer–Emmett–Teller (BET) surface area
Here,weaddressthequestionofwhetherenvironmentalsalinity
measurements. The BET surface area measurements were
alters biological responses of tilapia to CuO NP. The biological
carriedoutusingN gasadsorptioninanAUTOSORB-1usingan
responsesselectedinthisstudyasexperimentalendpointsinclude 2
optimized protocol (Quantachrome Instruments, Boynton Beach,
sublethal markers of toxicity such as behavioral stress and
FL,USA).Thesamplesweredegassedat120uCfor12hpriorto
opercular ventilation rate (OVR), which have been widely used
theadsorption-desorption cycle.
to monitor sublethal effects of diverse environmental stressors,
All data of NP characterization (TEM, DLS, ZP and ICP-MS
including toxicants, onfish [47,48,49].
can befoundinTable S1inFileS1).
Methods
Fish acclimation and care
CuO NP synthesis Tilapia maintenance and the experimental conditions were
subject to approval by the UC Davis Institutional Animal Care
CuO NP were synthesized using a flame spray pyrolysis
and Use Committee, under protocols #16604 and #16609.
techniquewithforcedjetatomizer(FigureS1inFileS1)similar
Throughout experiments, all efforts were made to minimize
tothemethodreportedpreviously[50,51].Thistechniqueiswell-
animalsuffering.Laboratory-raisedMozambiquetilapia(Oreochro-
suited forgeneratingenvironmentallyrelevant metaloxide NPin
mis mossambicus) were maintained in large closed circulation
high concentrations with control on particle size [52]. Copper
aquaria using de-chlorinated tap water (pH8.25, total hardness
nitratesaltwasdissolvedinethanoltoproducea0.3mMsolution
155.5 mg?L21) at 26uC61uC and 12h:12 h photoperiod. Two
foruseastheliquidprecursor.Thesolutionwassprayedintoa1:1
weekspriortoexposure,twelveanimals(10–12monthsold)were
hydrogen-nitrogenflameat2.0mL?min21withasyringepump.A
transferredtoexperimentaltankswith16Loffreshwaterforpre-
flow of argon was used as sheath at a rate of 10 L?min21. The
acclimation, and kept under those conditions until treatment was
precursor droplets were pyrolyzed to form NP in the high
started. Fish were fed daily with trout pellet diet (Silvercup SCD
temperatureflame.Particlesinthepost-flamegaseswereextracted
2.0 mm) at approximately 1% of fish body weight. For all
using a vacuum pump and collected in a baghouse (Filter
treatments,10%ofthewaterwaschangedeverydayforalltanks.
Specialists Inc; Part # BNM05P4PA). The collected powder was
Forthis,1.6 Lofthetankwaterwasremoved,and1.6Lofwater
washed several times with ultrapure water to remove any
supplemented with appropriate NP concentration and salt as
unreactedprecursorfromtheNP.TheNPweredriedforfurther
required(seebelow)wasadded.Thesalinitywasmonitoredwitha
use.
refractometerbeforeandafterthedailywaterchange.Agroupof
tanks (denoted as FW) were kept at freshwater conditions (tap
PLOSONE | www.plosone.org 2 February2014 | Volume 9 | Issue 2 | e88723
CuONPEffectsonFishatDifferentSalinities
water, with a salinity of 0–0.1 parts per thousand, ppt, Tissue samples were diluted with MilliQ water to an acid
corresponding to g?L21). On the other hand, tap water concentration of6%.
supplemented with sea salt (Instant Ocean aquarium seawater The concentration of elemental copper was quantified using
salt,Vancouver,USA)wasusedforasecondgroupoftanks(SW), ICP-MS at the Interdisciplinary Center for Plasma Mass
atamountsrequiredtodailyincreasesalinityby7 ppt/day,except Spectrometry at UC Davis. The instrument level of detection
for the first day when the change was from 0 to 3 ppt. For both (LOD) was 0.019 parts per trillion and the BEC (background
FWandSWgroups,animalswereexposedto0,0.5and5 mg?L21 equivalent concentrations) were 0.125 partsper trillion forCu.
of CuO NP through addition of a concentrated CuO stock ForstatisticalanalysisofCucontentintissues,data(inng?mg21
solution (10g?L21 in MilliQ water) that was probe-sonicated dry weight) were transformed to logarithms since the variances
(Branson Digital Sonifier 250, Branson Ultrasonics Corporation) provednottobehomogeneousasindicatedbyLevene’stest.The
with the maximum output of 200W, and with a frequency of transformed data were then analyzed by one way ANOVA
20kHz,for5 minutesbeforeadditiontothetanks.Toavoidany followed by theTukeypost-hoc test.
temperatureriseinthesolution,thesamplewassonicatedincycles
of 20s at 10% of maximum power, followed by 20s rest. After Enzyme activity assays
water changes, manual stirring was performed for 20seconds to Tissues were homogenized in 1.5-mL microcentrifuge tubes
ensure dispersion of the NP in the tanks. The salinity was after the addition of 5mL of homogenization buffer (50 mM
confirmed with a refractometer before and after the daily water potassium phosphate pH 7.4, 0.9% NaCl, 0.1% glucose, 1X
change. FW0 treatment is the handling control, where the fish proteaseinhibitorcocktailcOmpletMini,Roche)permgoftissue.
weresubjectedtothesameinteractionthantheanimalsintherest Homogenateswerecentrifugedat14000rpmand4uCfor20min;
of the tanks (water change, manual stirring, observation), but the supernatant protein concentration was determined by BCA
without salinity increase norNP addition. assay(Pierce,USA).Catalase(CAT),superoxidedismutase(SOD)
The animals were observed for signs of stress on a daily basis. and glutathione reductase (GR) enzymatic activity were deter-
The experimental exposure endpoint was established as the time mined using specific assay kits (Cayman Chemical, USA),
when more than 50% of the animals in the tanks showed following vendor instructions. Data were analyzed by One way
behavioral signs of stress (e.g., buoyancy difficulties, swimming ANOVA followed by Tukey’s post-hoc test to compare all the
abnormalities, lack of feeding) or when the mean opercular means andevery treatment pairseparatedly.
ventilationrate(OVR)was25%higherthantotheuntreatedFW
controlfortwoconsecutivedays.Theanimalsweresacrificed;the Oxidative status (GSH/GSSH ratio)
gill arches and liver were collected. The gill arches were rinsed Theoxidized(GSSG)andtotal(GSH+GSSG)glutathionelevels
withcoldPBSandgillepitheliumwasgentlyscrapedofffromthe were determined by the reaction with Ellman’s reagent 5,59-
cartilageandcollected.Tissueswerefrozeninliquidnitrogenand dithio-bis-2-nitrobenzoic acid, DTNB [54]. Tissues were homog-
stored at 280uCuntil used. enized in 10 volumes of 5% sulfosalicylic acid (SSA) or 5% SSA
Additionally, we performed a similar experiment in were all plus3 mMof1-Methyl-2-vinylpyridiniumtriflate(M2VP,Sigma),
experimental exposure conditions were replicated (aeration, fortotaloroxidizedglutathionerespectively[55].After30minof
stirring, water changes), except that no fish were included in the incubation on ice, samples were centrifuged (14000rpm, 4uC,
tanks. 20min),andsupernatanttransferredtonewtubes.Sampleswere
neutralizedtopH 7–7.2byadditionofneutralizationbuffer(2.125
Cu concentration determination by ICP-MS M Na CO ) and combined with assay mixture: 100mM
2 3
CuONPsamplesfromexperimentaltankswithorwithoutfish potassium phosphate pH 7.0 and 1mM EDTA, 0.13U?mL21
(takenbeforeofthedailywaterchangeexceptfortime=0,where glutathione reductase (Sigma), 0.04mg?mL21 NADPH (Fisher)
samplesweretakenimmediatelyafteradditionofNPtothetank) and0.0325 mg?mL21DTNB(Sigma).Absorbanceat405nmwas
were used to determine total Cu content and to measure soluble determined at 25minutes using a Spectrafluor Plus microtiter
copper.Totakerepresentativesamplesfromthetanks,westirred platereader(Tecan).OxidizedGSSG(Sigma)wasusedtobuilda
the tanks for 20sec before taking the sample. Additionally, each standard curve. Data were analyzed by One way ANOVA
sample (10 ml final volume) was taken by pipetting 56(2ml) followed by Tukey’s post-hoc test.
aliquots,whichwerecombined,fromdifferentpointsineachtank.
Samples were first digested with 70% HNO (trace metal grade Quantitative PCR (qPCR)
3
concentrated nitric acid, Optima, Fisher Scientific) at 70uC for Total RNA was extracted from tissues using TRIzol reagent
2 h. Then the samples were diluted to 6% HNO3 with MilliQ (Invitrogen, Life Technologies), and subsequently treated with
water and analyzed by inductively coupled plasma mass TURBO DNase (Ambion, Life Technologies). cDNA was
spectrometry(ICP-MS)fordeterminationofthetotalCucontent. produced from 1 mg of RNA using Random Hexamers and
To determine theconcentration of soluble copper ions, CuO NP ImProm II reverse transcriptase (Promega). Reactions were
samples from experimental tanks were loaded in dialysis bags performed in MicroAmp Fast Optical 96-well plate thermal
(MWCO 3,500Da, Spectrum Spectra/Por 3 RC) and dialyzed cycling plates (Applied Biosystems, Life Technologies) using Fast
overnight against water used for the fish tanks. This method has SYBRGreenMasterMix(AppliedBiosystems,LifeTechnologies)
beenshowntobesuitabletoseparatesolubilizedCu2+ionsandCu andprimerpairsdesignedwithPrimer-BLAST[56]:Metallothionein
asparticulatephase[53].Thedialyzedwaterwasdigestedin70% (59-GCGAGTGCGCCAAGACTGGAA-39 and 59-CAGCCG-
HNO anddiluted to6%HNO similar toother samples. GATGGGCAGCAGTC-39), Cytochrome P4501A (59-TGTCACC-
3 3
Tissue samples were treated in 70% HNO (Optima, Fisher GAGCACTATGCCACCT-39 and 59-TCCAGCTTTCTGTC-
3
Scientific) and hydrogen peroxide (34–37%, Technical, Fisher CTCGCAGTG-39), and b-actin (59-CCACAGCCGAGAGG-
Scientific)forelementalanalysis.Atfirst,wettissuesweredriedby GAAAT-39) and (59-CCCATCTCCTGCTCGAAGT-39). Reac-
overnight incubation in an oven at 110uC. Dry weight was tions were performed in a 7500 Fast Real-Time PCR system
recordedandtissueswerethendigestedin70%HNO at80uCfor (Applied Biosystems), using the following cycling conditions: 1
3
4 h, followed by hydrogen peroxide digestion overnight (70uC). cycleat95uC/20 sfollowedby40cyclesof95uC/3 sand60uC/
PLOSONE | www.plosone.org 3 February2014 | Volume 9 | Issue 2 | e88723
CuONPEffectsonFishatDifferentSalinities
30s. The Pfaffl method [57] was used to calculate the ratio of
mRNA expression in treated fish relative to the fresh water
handling controls, using b-actin as a reference transcript for data
normalization.DatawerecorrectedforindividualPCRefficiencies
(LinRegPCR software [58]). Data were then analyzed using
ANOVA on ranks (Kruskal-Wallis’s test) and Wilcoxon signed
rank statistic, giventhat these datawere non-parametric.
Results
CuO NP are highly crystalline and monoclinic
The flame-synthesized NP were characterized by TEM.
Figure 1A shows typical TEM micrographs of the CuO NP.
Quantitative analysis of these micrographs shows that the
synthesized CuO NP were polydisperse with a mean size of
21.2 nmandastandarddeviationof11.8 nm(n=245);77.4%of
the NP were smaller than 25 nm (Figure 1A, insert). The
darker appearance of some NP is mainly attributed to different
NP thickness and orientation when dried on the TEM grid, for
which electrons scatter incoherently and produce differential
contrast. The crystallinity and phase of the NP were determined
by powder X-Ray diffraction. The XRD pattern (Figure 1B)
confirms that the NP are highly crystalline and consist of
monoclinicCuO(JCPDS04-006-4186).TheXRDpatternofNP
afterprobesonicationwasalsoanalyzed(FigureS2inFileS1)
to investigate any possible physicochemical changes due to
sonication. The XRD patterns of the NP before and after
sonication are identical which confirms retention of their same
phase and crystallinity.
The surface area of the CuO NP was found to be
141.2 m2?g21 (the isotherm is presented in Figure 2A).
Figure 2B and Figure S3 in File S1 show the Zeta Potential
(ZP) of the NP samples taken from fresh water tanks over an 8
dayperiod.ThemagnitudeoftheZPcanbetakenasoneofthe
parameters to understand the colloidal stability of the NP [59].
NP with ZP values greater than +30 mV or less then 230 mV
typically have a high degree of stability in suspension [59]. We
analyzed the ZP of the NP dispersed in fresh water only
(Figure 2B); the ZP was seen to remain constant over time,
which indicates good colloidal stability of the NP. However, it
wasnotpossibletoobtaintheZPofNPdispersedinsalinewater
(i.e. at dissolved salt concentration ranging from 0.04–0.6 M
NaCl). As per the DLVO theory (Derjaguin, Landau, Verwey
and Overbeek), the agglomeration and stability of particle
dispersions are determined by the sum of the attractive and
repulsiveforcesamongindividualparticles.Theattractionwithin
the particles is mainly due to the van der Waals force while the
electrical double layer surrounding each particle offers electro-
staticrepulsiveforceamongparticles.Thezetapotentialandthe
thickness of the electrical double layer are two important
Figure 1. Characterization of CuO NP used in the present
parameters of the electrical double layer and an increase in study. (A) Representative transmission electron microscopy (TEM)
either will result in an increase in the electrostatic repulsive images of the CuO NP at two magnifications. The polydisperse size
interaction. The thickness of electrical double layer is a function distributionofCuONP(n=245)determinedfromTEMimagesisshown
of solution ionic strength, with an increase in ionic strength intheinsert.TheaveragesizeoftheNPisshown(mean6SD).(B)X-Ray
diffraction (XRD) pattern of the NP. The X, and Y axes of the XRD
leading to a decrease in double layer thickness. For ZP
representstheangles(2h)ofincidentX-raybeamandthecorrespond-
measurement, at a salt concentration of 0.001 M the dispersions
ingdiffractionpeakintensity.
aregenerally stable as the electrostatic repulsiveforces dominate doi:10.1371/journal.pone.0088723.g001
overtheattractiveforces.However,atanincreasedconcentration
of 0.01 M NaCl or higher, the attractive forces among the
solutions at these salt concentrations resulted in visible particle
particles started to dominate over the repulsive forces (due to
agglomeration and erroneous results. This observation is consis-
compression of the electrical double layer) resulting an unstable,
tentwithpreviousstudiessuchasthosefromJiangetal.[60]and
highly agglomerated dispersion. The concentration we used in
Brant et al. [61].
our study was 0.04 M at day 1 and stepwise increased to 0.6 M
NaCl at day 8. Therefore, any attempt to measure ZP of the
PLOSONE | www.plosone.org 4 February2014 | Volume 9 | Issue 2 | e88723
CuONPEffectsonFishatDifferentSalinities
volumeofwater),evenatthebeginningoftheexperiment(T=0),
particularly for SW tanks (Figure 4A). Within the first day of
exposure(T=1),thesuspendedCuONPconcentrationdecreased
from71%to52%inalltreatmentscomparedtotheCuOlevelsat
T=0 (Figure 4B). This trend was sustained over time. For
instance, the suspended CuO NP concentrations in SW0.5 and
SW5were73.4%and65.8%atT=3,54.7%and55.9%atT=6,
and 55.4% and 38.9% at T=8, respectively (Figure 4B, black
symbols). A comparable change in NP concentration over time
was observed in absence of fish (Figure 4B, white symbols). No
additionaldissolved(soluble)Cuionsweredetectedafterdialysisof
samples taken from any of the NP experimental tanks (data not
shown) when compared to the water from tanks without NP.
Therefore, the change in NP concentration might be due to NP
aggregation,oradsorptiontoglasssurfacesinthetanks.However,
visual inspection of the tanks did not reveal any sedimented
material, which can be attributed to the formation of sedimenta-
tion particles with a size not large enough to be visible by the
naked eye.
The state of aggregation of the CuO NP that remained
suspended during exposure was determined by measuring the
hydrodynamic size of the particles by DLS (Figure 4C). The
hydrodynamic size of suspended CuO NP was larger than that
observedbyTEM,rangingfrom93.54nm68.52inFW5atday5
to269.19 nm614.7inSW0.5atday6.ThetimecourseofCuO
NPhydrodynamicsizewassimilarforbothconcentrationsinFW
and for both concentrations in SW, as can be seen by the trend
linesinFigure4C.InFW0.5,FW5,andSW5thehydrodynamic
size fluctuated within a narrow range. In SW0.5, however, there
was a significant increase in hydrodynamic size at T=6 (38ppt,
corresponding approximately to ocean salinity), suggesting that
someCuONPaggregationhadoccurredunderthoseconditions.
ThiseffectmaybeduetoabridgingofNPviaionicbondstoform
Figure2.BETandZetaPotentialofCuONP.(A)BETisothermof NP-M+-NP (where M+ is the salt cation, Na+ in the case of
theCuONP(powder).(B)ZetaPotential(ZP)oftheNPsampletaken seawaterthatpromotesNPaggregation[62].Inaddition,divalent
fromFW5tanksoverthetime.Day0representstheZPoftheNPright cationssuchasMg2+andCa2+(presentintheseawatersalt)have
after mixing to the fresh water of the fish tank. The ZP of the NP
shown to induce aggregation of NP [63]. The concentration of
remainedfairlyconstantthoughouttheexperiments.Thedottedline
showedasaguide. monovalentcations(Na+-K+)inseasaltisapproximately8times
doi:10.1371/journal.pone.0088723.g002 higherthan Mg2+ andCa2+concentrations combined[64].
CuO NP exposure affects salinity-stressed fish differently CuO NP exposure alters the activity of antioxidant
than FW controls enzymes in liver and gills
Figure 3A shows salinity levels in tanks with a control fresh- TodeterminewhetherCuONPcantriggeranoxidativestress
waterenvironmentexposedfishandanimalsthatwerechallenged response in fish tissues, we measured the activity of three
with stepwise increasing salinity. Fish from both groups (salinity antioxidant enzymes in gills and liver after CuO NP exposure of
challenge -SW- and freshwater -FW-) were exposed to 0, 0.5 or fish(Figure5).Inbothliverandgills,CAT(Figure5A)andGR
5 mg?L21CuONP.Opercularventilationrate(OVR)meansfor (Figure 5C) showed significant activity changes, whereas no
each experimental group ranged from 62 to 84 beats/min changes in the total SOD activity were observed in response to
(Figure 3B). An increase greater than 25% in OVR (compared CuO NP exposure (Figure 5B). Liver CAT activity decreased
to FW0 control) was observed after 6 days of exposure to significantly in SW5 relative to FW0 controls. In contrast, CAT
increasing salinity (at 31ppt), in SW5 (26.0%63.3%). This activityincreasedsignificantlyinFW5comparedtoFW0controls
indicatingthattheresponsetoCuONPiscontext-dependentand
increase was sustained at day 7 (salinity=38ppt) and day 8
modulated by environmental salinity. GR activity increased
(salinity=45ppt) (Figure 3B, dotted line). Therefore, the
significantly at FW0.5 compared to FW0 in liver, and decreased
exposurewasstoppedatday8.Byday8,themeanOVRwasalso
at both FW5 and SW5 compared to FW0 in gills. Additionally,
significantly increased in all the other NP treatments, including
gills CAT activity decreased significantly in response to salinity
FW0.5 and FW5. In addition, the behavioral signs of stress
challenge and CuO exposure but there was no additive effect of
observedinthetreatmentgroupsweremildandpertainedtoonly
bothtreatments (Figure 5C).
aminorityoffishinaparticulartreatmentgroupatanygiventime
Another marker for oxidative stress in response to toxicants is
(Table1).
theamountofglutathione andtheratioofreduced(GSH) versus
oxidized (GSSG) glutathione [65,66]. Therefore, total
CuO NP concentration and size at different salinities
(GSH+GSSG) and oxidized (GSSG) glutathione concentration
NPconcentrationinwatersamplestakenfromeachgroupwas
were measured with the intent to estimate GSH levels and
below the expected value (based on mass added to a known
calculatetheratioGSH/GSSGingillsandliveroffishexposedto
PLOSONE | www.plosone.org 5 February2014 | Volume 9 | Issue 2 | e88723
CuONPEffectsonFishatDifferentSalinities
Figure3.ExperimentalconditionsduringCuONPexposure.(A)SalinitychangesregimeforFWandSWtanks.TheFWtankswerekeptin
constantfreshwater(0.1ppt,graytriangles),whileinSWtanks,salinitywasincreasedinastepwisemannereveryday(blacksquares),byadditionof
increasingamountsof seasalt.Salinityincreasestepswere of7ppt/day,exceptfor the firstdaywhenthe increasewasof 3ppt. (B) Opercular
ventilationrate(OVR)inexperimentaltanks,measuredasopercularbeats/minuteandexpressedasthedailymeanpertank(n=7).Thedottedline
representthe25%incrementselectedasexperimentalendpoint.AsterisksdenotesignificantdifferencesrelativetotheFW0perday(P,0.05).Top:
FW0,freshwater;FW0.5,FWplus0.5mg?L21CuO;FW5,FWplus5mg?L21CuO;Bottom:SW0,increasingsalinity;SW0.5,SWplus0.5mg?L21CuO;
SW5,SWplus5mg?L21CuO.n6tank=7.
doi:10.1371/journal.pone.0088723.g003
CuO NP (Figure 6). Total GSH+GSSG levels increased proteins, is involved in detoxification of many toxic compounds
significantly in SW0 liver compared to FW0 but in none of the [68]andisencodedbyametal-responsivegeneinfish[69,70,71].
other treatments (correlated with an increase in GSH levels), The CYP1A promoter region in fish contains putative metal-
although all SW groups showed the same trend. The only responseelements,MREs[72]andMTinductionhasbeenshown
significant change with respect to the GSH/GSSG ratio was a tooccurafterionicCuexposure[73].mRNAabundanceswerein
decrease intheratioin gillsfollowing theFW0.5treatment. generalstronglyaffectedbyCuONPinbothliverandgills,when
comparedtotheFW0control(Figure7,asterisks).Forexample,
Metal-responsivegenesareinducedbyCuONPenzymes CYP1Aexpressionwasgreatlyinducedinliveraftertreatmentwith
We analyzed the expression of two genes involved in 0.5 mg?L21 CuO NP in both FW and SW (1104% and 738%,
detoxification against metals, such as Metallothionein (MT) and respectively),anditwasalsoslightlyinducedinFW5(Figure7A).
Cytochrome P450 1A (CYP1A). MTs are ubiquitous, low-molecular In gills, CYP1A was increased to 238% in FW0.5 and 276% in
weight, Cys-rich proteins, capable of binding heavy metals SW5, whereas it was repressed to 50.1% in FW5. Salinity alone
(including Cu), which typically show an up-regulated gene did not affect CYP1A expression as it can be seen in SW0
expression when intracellular metal concentrations increases treatmentinlivercomparedtoFW0,althoughingillsweobserved
[67]. CYP1A, belonging to the Cytochrome P450 family of
PLOSONE | www.plosone.org 6 February2014 | Volume 9 | Issue 2 | e88723
CuONPEffectsonFishatDifferentSalinities
conditions. The osmotic stress induced by exposing fish to
Table1. Stress signsonfish observed during the CuONPs
environmental salinity fluctuations has been well-characterized,
exposure.
from sensing mechanisms in fish [74] to the study of strategies
developed by the organism to cope and adapt to the osmotic
changes [75,76]. Therefore, the use of hyperosmotic stress in
Tank Day Stresssign n % Intensity
tilapiaitsausefultooltounderstandhowNPmayhaveeffectson
FW0.5 6 Lossofbuoyancy/swimmingcontrol 2 16.7 Mild
fish under different environmental conditions. In laboratory
FW5 5 Lateresponsetofeeding 4 33.3 Mild conditions, gradual (stepwise) acclimation to salinity changes has
6 Lossofbuoyancy/swimmingcontrol 1 8.3 Mild beenlargelyusedtomodelosmoticstressindifferentfishspecies,
7 Lateresponsetofeeding 5 41.7 Mild allowing a better assessment of chronic effects of osmotic stress
[77].Salinitychangesintheenvironmentcanbeseasonal,diurnal
SW0 5 Lossofbuoyancy/swimmingcontrol 1 8.3 Mild
and spatial, depending on factors such as rainfall, tides, runoffs
SW0.5 5 Lossofbuoyancy/swimmingcontrol 2 16.7 Mild
andevaporation[78].Severaltilapiaspecieshavebeendescribed
SW5 5 Lossofbuoyancy/swimmingcontrol 2 16.7 Mild
in FW to hypersaline habitat conditions [41]. Their high salinity
7 Lateresponsetofeeding 4 33.3 Mild tolerancemayhaveevolvedasaresultofstrongselectionpressure
associated with frequent seasonal droughts and intermittent
n:numberofanimalspertankshowingthestresssign(thepercentageof
flooding events in areas characterized by salt-rich bedrock and
stressedanimalsisalsoshown).Intensity:representsatwhatextentthe
observedstresssignsaffectedthenormalbehavioroftheanimals. soil[79].Forexample,thehybridspeciesCaliforniaMozambique
doi:10.1371/journal.pone.0088723.t001 tilapialivesin44pptattheSaltonSea,southernCalifornia[80],
while another tilapia species (black-chinned tilapia, Sarotherodon
anincreasedexpressionofCYP1AinSW5whencomparedtoFW5 melanotheron) can tolerate hypersaline environments (,130ppt) in
(1 symbol). an african saline estuary [81]. Species in estuaries and tide pools
In liver MT (Figure 7B) showed a 379% increase in FW5, (which include tilapia species) are exposed to short term salinity
while a reduction in MT transcript levels was observed in SW0 fluctuations, and even to vertical or horizontal salinity gradients
(12.8%) and SW0.5 (17%) when compared to FW0. In gills, MT [76]. For example, the estuarine fish Fundulus heteroclitus may be
transcript accumulatedto204%and209%forFW0.5andFW5, exposedtodailyfluctuationsinsalinityrangingfrom,0.1 pptto
correspondingly, while it was reduced to 49.9% in SW0, always ,35ppt[82].
relative to FW0. A significant decrease of MT expression in liver In our study, sub-lethal signs of behavioral and physiological
was observed in SW-NP treatments when compared to the FW distress, rather than mortality, were chosen to define maximal
treatments(1symbol),althoughinSW5thelevelsarecomparable exposurelevels.TheNPconcentrationsusedinthepresentstudy
to FW0 (so increased relative to SW0 and SW0.5). Therefore, were chosen based on previously reported experiments, where
CYP1AlevelsareingeneralhigherintheliverinresponsetoCuO CuO NP exposure to fish caused toxic effects (see Introduction).
NP exposure. In contrast, MT was generally induced in FW Throughout all exposures, only mild signs of distress were
treatmentsandrepressedinmostSWtreatments(whencompared
observed at the whole organism level, and consistently apparent
toFW0)inbothliverandgillssuggestingamodulationofCuONP
inalowproportionofthefishinanytank(Table1).Weobserved
effects by environmental salinity.
a sustained OVR increase for group SW5, prompting us to stop
exposureday8(Figure3B).Asobserved,theOVRwasaffected
Copper content in tissues exposed to CuO NP
in a CuO NP concentration-dependent manner in SW, but the
CuconcentrationwasdeterminedbyICP-MSinliverandgills responseinFWdidnotshowsuchatrend(althoughbyday8,we
of animals exposed to CuO NP (Figure 8). In both tissues there observed an increased OVRat bothNP concentrationsinFW).
was a consistent trend towards an increase in Cu content with RegulationofOVRisasympatheticresponsetostress[83]that
increasedconcentrationofCuNPintheenvironment.Thistrend canservetocompensateforareductioninpartialO pressurein
2
wasalsotrue fortheliver,althoughthesignificancethresholdfor bloodoffish[84,85,86].EngineeredNPhavebeenshowntoaffect
FW treatments was not exceeded in any case when CuO the respiratory system in mammals and, analogously, the gill
treatments were compared to the FW0. However, the content of functionandstructureinfishes[45,87,88].Moreover,anincrease
CuinSW0wasinferiorwhencomparedtotheFW0andtoFW5,
in ventilation rate and respiratory disturbances have been
SW0.5 and SW5, suggesting an accumulation of Cu in SW
associated with exposure to NP [89,90]. These earlier findings
treatments (when compared to the SW0 treatment). In gills, Cu
areconsistentwithourobservationthatanincreaseinOVRisfirst
accumulatedsignificantlyinbothSWtreatments,butnotinFWat
observed at thehighestconcentrationof NPincombination with
0.5 mg?L21 CuO when compared to FW0. Additionally, accu-
additive salinity stress. Interestingly, one common response
mulation of Cu in SW0.5 was significantly different to FW0,
observed in gills after exposure to different nanomaterials was
FW0.5 and SW0, while FW5 and SW5 treatments were higher
theproductionofmucusintroutandzebrafish[87,90,91].Mucus
whencomparedtotheirrespectivenonNPtreatment,althoughit
productionhasbeenreportedasaresponsetomanytoxicantsfor
was comparable between them.
providing a barrier that prevents toxin interaction with epithelial
ThesedatashowthatevenwhenCuiseffectivelyaccumulated
cells [92,93]. For instance, carbon nanotubes cause respiratory
in both tissues when fish are exposed to CuO, the degree of this
toxicity, increased mucus production, and gill damage in trout
accumulation is to some extent dependent on the environmental
[90]. Thus, in addition to inducing oxidative stress, NP may
salinity, since in SW treatments, Cu accumulated effectively
thickenthemucuslayerandhindergasexchange,whichmightbe
followed NP treatment comparedtotheSW0 inbothtissues.
compensated by elevating OVR. Several studies suggest that the
production and composition of gill mucus depends on environ-
Discussion
mental salinity [94,95]. Our finding that OVR only increases
Salinity changes represent a relevant variable to test if NP greatlyinSW5,ratherthaninFW5,supportsthenotionthatgill
toxicity in fish is affected by exposure to different environmental mucus production depends on environmental salinity and
PLOSONE | www.plosone.org 7 February2014 | Volume 9 | Issue 2 | e88723
CuONPEffectsonFishatDifferentSalinities
Figure4.CuONPbehaviorinexperimentaltanks.(A)CuONPconcentrationduringexposureinwatersamplesfromtheexperimentaltanks
(showedmean6SD).(B)TimecourseofCuONPconcentration(mean%6SD)relativetoT=0forinFW0.5andFW5(left),SW0.5andSW5(right)
treatmentgroupswithfish(+F,blackcircles)andwithoutfish(2F,opencircles).Salinity(inppt)ateachtimepointisdepictedforSWtreatments.(C)
CuONPsizeinexperimentalFW(left)andSW(right)tanksanalyzedbyDLS.Theinsetdepictsanexampleofthelog-normaldistributiondataforthe
FW0.5tankatdays0and6.Lineartrendlinesforeachtreatmentareshown.Additionally,tanksalinityinrelationtotimeofexposureisshown(light
graylines).FW0,freshwater;FW0.5,FWplus0.5mg?L21CuO;FW5,FWplus5mg?L21CuO;SW0,increasingsalinity;SW0.5,SWplus0.5mg?L21CuO;
SW5,SWplus5mg?L21CuO.T=0correspondstothesampletakenimmediatelyafterCuONPadditionandtheinitialsalinityincrease.T=#:sample
takenafter#daysofCuOexposure.
doi:10.1371/journal.pone.0088723.g004
PLOSONE | www.plosone.org 8 February2014 | Volume 9 | Issue 2 | e88723
CuONPEffectsonFishatDifferentSalinities
Figure5.Analysisofenzymaticactivity.(A)Catalase,(B)SuperoxideDismutaseand(C)GlutathioneReductaseweremeasuredinhomogenates
from liver (black bars) and gills (white bars) after exposure to CuO NP in different environmental salinities. FW0, fresh water; FW0.5, FW plus
0.5mg?L21CuO;FW5,FWplus5mg?L21CuO;SW0,increasingsalinity;SW0.5,SWplus0.5mg?L21CuO;SW5,SWplus5mg?L21CuO.Activityis
expressedasmeannmol.ml21.min21.mgprot216SEM.GroupswithsignificantdifferentmeansbyAnovaanalysisareshownwith`(Pvalue).Letters
denotegroupsshowingnonsignificantdifferencesbyTukey’spost-test(P,0.05).n6treatment=5.
doi:10.1371/journal.pone.0088723.g005
illustrates the importance of the environmental context for were more marked in SW over (i) the OVR; (ii) intracellular Cu
assessing theimpactof NP. accumulation (higher in both liver and gills in SW0.5 and SW5
The effective CuO concentration measured in the tanks was compared to SW0 treatment, while in FW the accumulation was
substantially lower over time for all treatments, in particular for observedonlyatthehigherNPconcentration,Figure8);and(iii)
SWtanks(Figure4A).However,therelativeCuONPdecreased metal-responsive gene expression. These effects are mediated by
in a similar manner both in presence or absence of fish theNPstillsuspended,whichdidnotpresentamarkedincreasein
(Figure 4B), showing that the decrease in the NP concentration their hydrodinamic size. This represents an interesting scenario,
overtimeismorelikelyduetosedimentationand/oradsorptionto where environmental conditions affects effective NP concentra-
tank surfaces. This effect is more evident in SW than FW, tion, although the toxic effects of NP are not decreased, but
suggestingthatthesedimentationeffectmightbemediatedbythe enhanced,supportingthenotionthatNPtoxicityisdependentin
increasinglevelsofmonovalentanddivalentcationswhilesalinity theenvironmentalconditions.ThemeasuredreductioninNP(Cu)
was increased during exposure. However, it is important to concentration in the tanks cannot be accounted for by the
emphasize that despite ofthe decrease intheeffective concentra- accumulation in fish – the loss of Cu and NPs can only be
tiontowhichthefishwereexposedparticularlyinSW,theeffects explained by the unavoidable loss of particles to surfaces in the
PLOSONE | www.plosone.org 9 February2014 | Volume 9 | Issue 2 | e88723
CuONPEffectsonFishatDifferentSalinities
Figure6.GlutathionelevelsinliversandgillsofCuONPexposedfishindifferentsalinities.Totalglutathione(GSH+GSSG,toppanel);
Oxidizedglutathione(GSSG)andreducedglutathione(GSH)levels(centerpanel).TheratioGSH/GSSHisshowninbottompanel.FW0,freshwater;
FW0.5,FWplus0.5mg?L21CuO;FW5,FWplus5mg?L21CuO;SW0,increasingsalinity;SW0.5,SWplus0.5mg?L21CuO;SW5,SWplus5mg?L21CuO.
Expressedasmeannmol.mgtissue216SD.GroupswithsignificantdifferentmeansbyAnovaanalysisareshownwith`(Pvalue).Lettersdenote
groupsshowingnonsignificantdifferencesbyTukey’spost-test(P,0.05).n6treatment=5.
doi:10.1371/journal.pone.0088723.g006
tank that cannot be quantified directly in any aquatic toxicology uptake Ag and CeO NP [97]. In trout, hepatocytes can
2
study of thissort. accumulate CeO , TiO and Ag NP in vitro [98], while when
2 2
Someconsiderationscanbemadetoexplaintheaccumulation TiO NP are administered intravenously, they are deposited
2
ofintracellularCu.Ononehand,severalstudieshaveshownthat intracellularlyinkidney[99].AnotherpossibilityisthatNPcould
NPareinternalizedbycells(evenNPwithasizeof500nm).The becoatedbyfish-originatedorganiccompounds(e.g.,proteins)or
mechanismsinvolvedintheinternalizationcanbeactivetransport adsorbedinmucus[6,100],affectingthestabilityoftheNPsurface
(endocytosis) or passive diffusion through the plasma membrane in the vicinity of the exposed tissues, such as gills. Thus, a local
[13,96].Forexample,hepaticandintestinalhumancelllinescan release of ions is produced, which are uptaked by the proximal
PLOSONE | www.plosone.org 10 February2014 | Volume 9 | Issue 2 | e88723